Testing reference sensor with oscilloscope ... how?

[billjam]

I am having some problems getting consistent, logical signals from the flywheel reference sensor (standard 3.6 sensor) at the MS3X ECU and I need help to figure out how to check the sensor with an oscilloscope as per the Porsche manual. Then the sensor can either be ruled in or out as the problem.
I should add that I have not been able to get good signals in the TunerStudio tooth logger. Sometimes I get nothing, sometimes a partial signal, sometimes a full 60-2 signal.

I have connected the oscilloscope to the sensor and get a trace running across the screen, but that's where I come to a dead stop. When I crank the engine, the trace goes all over the place and I can't figure out how to tune it so that it looks like the diagram in the manual below. If you hadn't already guessed, I now nothing about oscilloscopes ... I'm amazed that I even got this far!

There's lots of oscilloscope guides and how to's on the net, but I haven't been able to get past this stage, so I need some help from someone who has done this or knows how to setup this oscilloscope.
Can someone who is familiar with this type of scope give me some simple steps to follow? Pics below show current settings.

[Tippy]

Have you tried capture the signal inside the ECU? That's where I checked mine with a handheld o-scope that autoscales.

Might be easier.

[billjam]

Ha! Might be easier for you!!!
You are dealing with a real simpleton here when it comes to some of this electronic stuff. I have only got as far as I have with bulk assistance from you and others here.

The scope leads are plugged into the sensor connector and I have the oscilloscope pictured above, so playing around inside the ECU with a different type of scope isn't really an option right now.
It is all hooked up ready to measure, I just don't know how to setup the scope to show meaningful info.

I have already tried a few tooth logs. Isn't that is effectively capturing the signal from inside the ECU? This signal is erratic, that's why I need to get back to the source so I can be sure the sensor is OK.

[MConn]

billjam,
Here's a picture of the speed and reference waveforms that i took from the ECU pins.



It's a lot of trouble to open the ECU to install the scope leads, but that way you are checking the wiring and sensors.

The sensor needs to be checked in-circuit because it's getting power from the ECU - you cannot unplug the sensor at the motor and connect the scope to the sensor.

Mike

[billjam]

Thanks Mike, but I didn't think that the 3.6 crank sensor was powered. I thought it just created its own pulses as the flywheel teeth passed by.
I can already check the output of sensor by running a tooth log in TunerStudio, but these logs are inconsistent, which is why I want to test the sensor directly without going through the ECU and wiring.
The way I read the manual instructions above, the sensor is being checked at the ECU connector with it disconnected from the ECU (otherwise you can't access the pins).

[Quicksilver]

+1 ^

The sensor creates a power pulse and is unpowered. Disconnect from DME to test. When I went all oscilloscope on mine I pulled the plug at the DME and checked it there. If it is good there then you know the DME is getting good signal.

[MConn]

I should have read your post closer. Sorry if I provided bad information. I must have cabin fever.

Mike

[scarceller]

Trying to look at the signal during cranking is very difficult with a simple scope, you need a storage capable scope for that. If however the car will start and idles you should be able to see the signal with any scope.

If it's a no start issue you can quickly sanity check for signal at cranking with a 12vdc LED test light but it it MUST be a LED type. Walmart sells one in the auto section of the store for like $5.00 simply put the light across both pins and crank the motor and the light should flicker, if not something is wrong.

[jpnovak]

Bill. The scope just has two connections. These will connect to the Reference sensor. Just put the ground lead on the sensor shield and the signal pin to the sensor. You can connect the sensor lead anywhere along the path between the sensor and the ECU. You might be able to pull back a connector and probe the crimped wire. Just find a place that is convenient.

You can crank the car with the scope connected. Change the Voltage scale such that the signal is on scale. Most likely you could start on a 25mV/div and come close. Then change the time scale such that you can see the full waveform. A good starting point would be 100ms scale.

You can watch the signal while cranking but as Sal said a storage scope is best.

[billjam]

I guess I owe you all an update. I haven't had time to look at the car for a couple of weeks, but managed to get home early today so I started from scratch again.
With the scope attached to the crank sensor's connector, with a bit of knob twirling it showed the following trace.

I have absolutely no idea what that represents, but when I cranked the engine, the trace changed to this ...

I don't know what that means either, but at least it showed that there was some action in the sensor circuit.
Bouyed by this, I moved the scope probes to the ECU end of the harness and got exactly the same trace.
So far so good!

Next step was to reconnect the ECU and try a tooth log again. Bugger me, it was a perfect log.


Then I tried test modes for spark and fuel and I could hear coils firing and injectors working exactly as they should.
Well the only step left now was to actually try and start the engine.

Sorry guys, it's getting late and I have an early start tomorrow, so I'll finish this tomorrow.

[jpnovak]

Bill. The first oscillating wave in your scope looks like AC line voltage. Are you set for AC or DC trigger? Should be DC trigger. This should give you a flat line. This will allow you to resolve the second imposed wave in your second picture.

Your ECU trigger log looks good.

Now compare with the car running. I wonder if you are picking up EMI noise from one of your systems that prevents the car from running right.

[scarceller]

What's the voltage scale set at in that first pic?
And what's the freq set at?
Would be helpful to know this.

Also is the first pic with the scope just connected? engine not turning at all? If so something is wrong, as Jamie pointed out you are most likely picking up 60 cycle noise?

[Quicksilver]

Originally Posted by billjam

Hard to tell from the picture but it looks like the AC/DC select is extended in the AC position.

[T77911S]

connect the ground connection on the O scope to the car ground. the body itself. its on the front of the scope next to ch2.


AC/DC on the ch should not matter. use AC though. that basically for setting a reference like if you wanted to measure a signal with reference to ground.

[billjam]

Thanks for your hints and suggestions guys.

Originally Posted by scarceller

What's the voltage scale set at in that first pic?
And what's the freq set at?
Would be helpful to know this.

Also is the first pic with the scope just connected? engine not turning at all? If so something is wrong, as Jamie pointed out you are most likely picking up 60 cycle noise?

1 volt and 10ms.
Yes, this is just with scope connected, engine not turning.

Before I got to the stage of rerunning the scope test of the crank sensor with different settings, I figured it was worth a try to see if the engine would start, as everything seemed in order ... except for my crappy scope operating skills!
Anyhow, it started up and ran fine, so whatever the original problem was, after all my fiddling with the innards of the ECU and disconnecting and reconnecting everything, I seemed to have fixed the problem.

Thanks again for your input, I guess we'll have to wait for another opportunity to turn me into a oscilloscope expert.

[T77911S]

probably a bad connection.

[rick-l]

do you have the scope probe on terminals 1 and 2 (ground) of the sensor disconnected from the car?

[billjam]

Yes Rick. I was measuring the sensor output directly at the disconnected sensor connector pins 1 and 2.

Car is running fine now, just need to address poor cold starting and inop IACV and passenger electric window and oil leak and rattling gear linkage and slow puncture and torn headliner and body repaint and, well you get the picture!

[fetus]

Are there any handheld oscilloscopes that can show the same thing?

[TimT]

Check out Seeed Studio



https://www.seeedstudio.com/DSO-Nano-v3.html